JP2006015431A - Robot controller and control method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To control the fingertip part of a robot so as to follow a set target track and to effectively reduce the deviation of an actual track of the robot with respect to the set target track. <P>SOLUTION: A CPU of the robot controller sets a halfway target point by subdividing the target track for each prescribed sampling time and outputs a positional command (an angle command) Pt so as to move the fingertip part of a robot body to the halfway target point. The position of the fingertip part of the robot body from the current time (k) to the time ahead by a prescribed sampling time (k+n) is estimated by using a model. A correction value for the positional command is calculated after obtaining the deviation between the estimated position and the target track. The current positional command Pt(k) is corrected by the calculated correction value. A perpendicular vector Pα drawn on the target track from the estimated position of the fingertip part of the robot body is made as the correction value. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a robot control apparatus and a control method for controlling a hand portion of a robot to follow a set target trajectory.

  For example, in an articulated (6-axis) type robot that performs operations such as processing, assembly, sealing, and welding, the position of the hand portion of the robot is made to follow a preset target trajectory by the robot controller. In addition, the servo motor of each axis is controlled. In general, a feedback control method is adopted for this control. However, in general feedback control, a response delay is inevitably caused in each servo motor, so that there is a problem that the actual trajectory of the robot deviates from the target trajectory.

Therefore, in recent years, in order to eliminate the deviation of the trajectory in the feedback control, as in Patent Document 1, an error between the teaching trajectory and the trajectory when the actual robot is operated is obtained. A technique for performing adjustment to correct a teaching point so as to be along a desired teaching locus is considered. Further, Patent Document 2 discloses a technique for adopting feedforward control in order to reduce a delay of a servo system when performing an interpolation operation by a robot, and controlling the position of the robot to always coincide with the command position. It is shown.
Japanese Patent Laid-Open No. 11-48176 JP-A-8-99278

However, since the technique described in Patent Document 1 requires adjustment by repetitive operation for each operation of the robot, adjustment is also required when the operation condition such as the operation speed is changed. Further, when the operation is changed due to an external condition or the like, there is a disadvantage that adjustment for all the operations is necessary, and the adjustment man-hour becomes enormous.
Also, the feedforward control as shown in Patent Document 2 has a complicated shape like an articulated robot, and its characteristics become nonlinear due to the influence of gravity, inertial force, load fluctuation, etc. It is difficult to make the deviation zero. Since the position of the robot is controlled so that it always matches the target position command, the position command may change rapidly. In this case, the robot vibrates in an attempt to force the position command to follow. After all, there is a problem that the effect of reducing the deviation of the orbit cannot be obtained sufficiently.

  The present invention has been made in view of the above circumstances, and its purpose is to control the robot's hand to follow the set target trajectory, and to offset the deviation of the actual robot trajectory from the target trajectory. It is an object of the present invention to provide a robot control apparatus and control method that can be effectively reduced.

  In the work to be performed while the robot's hand portion follows the set target trajectory, the time on the target trajectory is used, for example, when machining the workpiece with a processing tool, or for applying a sealant or welding. There is a case where it is not required to be so strict and it is more important to follow the target trajectory correctly. The present inventor does not always match the position command as in the feedforward control when controlling the hand portion of the robot to follow the set target trajectory. Focusing on the control to place the position of the part on the target trajectory, it is confirmed that this makes it possible to eliminate the sudden change of the position command and easily follow the target trajectory. The invention was accomplished.

That is, the robot control apparatus of the present invention controls the robot's hand to follow the set target trajectory, and the robot trajectory is subdivided into the target trajectory at each sampling time. Position command output means for outputting a position command so as to move the hand portion, a control portion for controlling each axis of the robot based on the position command, and a position of the hand portion of the robot after a predetermined sampling time from the present An estimation unit for estimation, a correction value calculation unit for calculating a correction value of the position command by obtaining a deviation between the position estimated by the estimation unit and the target trajectory, and a correction value calculated by the correction value calculation unit And a correction means for correcting the current position command (invention of claim 1).

  Further, the robot control method of the present invention outputs a position command so as to move the hand of the robot to a target point on the way of subdividing the set target trajectory at each sampling time, and based on the position command There is a method for controlling each axis of a robot so that the hand portion of the robot follows the target trajectory, and estimates the position of the hand portion of the robot ahead of a predetermined sampling time from the present time, and estimates The present invention is characterized in that a deviation between the calculated position and the target trajectory is obtained to calculate a correction value for the position command, and the current position command is corrected based on the calculated correction value (invention of claim 3).

  According to the robot control device and the control method of the present invention, when the robot's hand portion follows the target trajectory, the deviation of the robot position from the target trajectory after a predetermined sampling time from the current position of the robot is estimated. The current position command is corrected according to the estimated deviation. As a result, the position deviation of the robot, which is estimated to occur first, is controlled in advance by correcting the position command in front of it, and the position deviation from the target trajectory of the robot is reduced. Can be made. At this time, control is performed with a certain delay with respect to the target point on the way for each sampling time, but vibration due to excessive follow-up to the command value can be prevented, and deviation from the target trajectory is effective. Therefore, the actual position of the hand portion of the robot can be accurately followed.

  For the “predetermined sampling time ahead”, an appropriate time can be determined experimentally (or empirically) in advance in consideration of characteristics of the robot. Further, as a means for estimating the position of the hand portion of the robot, a well-known technique such as constructing an approximate model of the robot and solving the model for discretization can be used.

  And in this invention, when calculating the said correction value, it can comprise so that a correction value may be calculated from the estimated position of the hand part of a robot as a correction vector by using the perpendicular vector dropped on the target trajectory. (Inventions of Claims 2 and 4). According to this, the most efficient correction value can be obtained, and it becomes more effective.

  Hereinafter, an embodiment in which the present invention is applied to an articulated robot for assembly will be described with reference to the drawings. FIG. 5 schematically shows the external configuration of the robot system. This system includes a robot body 1 and a robot controller 2 as a control device according to the present embodiment for controlling the robot body 1, and a teaching pendant 3 is connected to the robot controller 2.

  The robot body 1 is configured as a 6-axis small vertical articulated robot for assembly in this case. Specifically, the robot body 1 includes a shoulder portion 5 on a base portion 4 so that the shoulder portion 5 can rotate (turn) around a vertical axis, and a pair of left and right end portions of the shoulder portion 5 extending upward in the drawing. The lower arm portion 6 is connected so as to be rotatable about a horizontal axis, and the front end portion of the intermediate arm portion 7 is rotated about the horizontal axis so as to be sandwiched by the upper end portion of the lower arm portion 6 from both sides. It is linked movably.

  Further, a rear end portion of the upper arm portion 8 extending forward is connected to the front end portion of the intermediate arm portion 7 so as to be coaxially rotatable, and the front end side of the upper arm portion 8 is bifurcated left and right. The wrist part 9 is connected so as to be rotatable about a horizontal axis. Further, a circular flange portion 10 is connected to the front end portion of the wrist portion 9 so as to be coaxially rotatable. Although not shown, a hand such as a workpiece gripping tool, a processing tool, a sealing material application head, and a welding head are attached to the flange portion 10. The center portion of the tool mounting surface of the flange portion 10 is a hand portion.

  In the robot body 1, a servo motor 11 (shown only in FIG. 4) that is a drive source for driving each axis, and a transmission mechanism such as a belt transmission mechanism and a speed reducer (not shown) are incorporated. . As shown in FIG. 4, each servo motor 11 is provided with a rotary encoder 12 as a position detection sensor. The servo motors 11 for these axes are controlled by the robot controller 2. In FIG. 4, the shoulder portion 5, the lower arm portion 6, the intermediate arm portion 7, the upper arm portion 8, and the wrist portion 9 are shown as one movable block, and only one servo motor 11 is represented. Show.

  On the other hand, as shown in FIG. 4, the robot controller 2 includes a CPU 13 as a control means, a ROM 14, a RAM 15 and an interface 16 connected to the CPU 13, and a drive as a drive means for driving the servo motors 11. A circuit (servo control unit) 17 and a position detection circuit 18 as position detection means are provided. The ROM 14 stores a control program for the entire robot, and the RAM 15 stores an operation program set based on teaching data by the teaching pendant 3 or the like. The teaching pendant 3 is connected to the interface 16 and peripheral devices such as an image processing device, a monitor device, and a personal computer (not shown) can be connected.

  The position detection circuit 18 detects the current position (rotation angle of each joint) of the servo motor 11 of each axis based on the pulse signal from the rotary encoder 12, and the detected position information is: The drive circuit 17 and the CPU 13 are provided. The drive circuit 17 compares the position command given from the CPU 13 with the current position (angle) given from the position detection circuit 18 and supplies the servo motor 11 with a current corresponding to the deviation to drive (feedback control). It functions as a control unit. Thereby, the position of the hand part of the robot body 1 (which is a concept including a three-dimensional position and orientation) is controlled so as to follow the target trajectory L (shown by a solid line in FIG. 3) set in the operation program. Thus, a predetermined work (assembly work or the like) is performed.

  Now, the robot controller 2 subdivides the target trajectory L every predetermined sampling time (for example, 8 ms) by a software configuration centered on the processing of the CPU 13, and indicates intermediate target points (indicated by black circles in FIG. 3). ) Is set, and a position command (angle command) Pt is output so as to move the hand portion of the robot body 1 to the target point on the way. At this time, in this embodiment, as will be described later in the description of the operation, the robot controller 2 (CPU 13) determines the position of the hand portion of the robot body 1 at a predetermined sampling time (k + n) from the current (k) (FIG. 3). The position command correction value is calculated by calculating the deviation between the estimated position and the target trajectory L, and the current position command Pt is corrected by the calculated correction value. ing.

  Therefore, the CPU 13 functions as a position command output unit, an estimation unit, a correction value calculation unit, and a correction unit. Further, in the present embodiment, as shown in FIG. 3, in calculating the correction value, the perpendicular vector Pα dropped on the target trajectory L from the estimated position of the hand portion of the robot body 1 is used as the correction value. It is like that. For the predetermined sampling time ahead, an appropriate time can be experimentally determined in advance, for example, 50 ms ahead in consideration of the characteristics of the robot body 1 and the like. As a means for estimating the position of the hand portion of the robot body 1, a known method of constructing an approximate model of the robot body 1 and solving the model to discretize it can be used.

  Next, the operation of the above configuration will be described with reference to FIGS. Now, as shown in FIG. 3, in the set operation program, assuming that the hand portion of the robot body 1 is controlled to follow the target trajectory L, sampling times (k, k + 1, k + 2,..., K + n, ..), A position command (Pt (k)) is output so that the hand of the robot body 1 comes to the target point on the way.

  However, when normal feedback control is performed, a response delay is inevitably caused in each servo motor 11, so that the actual trajectory of the hand portion of the robot body 1 is shifted from the target trajectory L as shown by a broken line in FIG. There are circumstances that cause a gap. In FIG. 3, white circles show examples of positions where the hand portion of the robot body 1 actually moves at each sampling time (k, k + 1, k + 2,..., K + n,...) When normal feedback control is performed. ing.

Therefore, in this embodiment, as shown in the control block diagram of FIG. 1, the position command (Pt) is corrected and the position control of the robot body 1 is performed.
That is, first, a position command (command angle Pt (k + n)) is input to the estimation model unit 19. In this case, the position command Pt (k + n) is output n samples ahead of the current (k) midpoint target point. In the estimation model unit 19, the position (joint angle) of each axis after n samples is estimated using the model, and the estimated position (angle) is input to the locus deviation estimation unit 20. The trajectory deviation estimation unit 20 obtains a deviation between the estimated position after n samples and the target trajectory L. In the locus correction vector calculation unit 21, a correction command value Pα is calculated. Here, as shown in FIG. 3, a perpendicular vector dropped from the estimated position n samples ahead onto the target trajectory L is used as the correction vector. The calculated correction value Pα is input to the adder 23.

  On the other hand, the position command (command angle Pt (k + n)) is input to the delay element 22, and the position command Pt (k) at the current sampling time (k) is output from the delay element. Then, the adder 23 adds the correction value Pα to the position command Pt (k), and the position command is set to Pt (k) + Pα and is input to the control loop 24 of the robot. Although detailed illustration is omitted, the control loop 24 of this robot is a loop for performing general feedback control.

  The flowchart of FIG. 2 shows the procedure for correcting the position command in the above-described control. That is, in step S1, a position command value n samples ahead from the present is calculated from the target trajectory L set in the operation program. In the next step S2, the position of the hand portion of the robot body 1 after n samples is estimated. In step S3, an estimated value (perpendicular vector) of the locus deviation is calculated, and in step S4, the current position command Pt (k) is corrected by the estimated value (correction value Pα) of the locus deviation. Is called.

  As a result, the positional deviation from the target trajectory L of the hand portion of the robot body 1 that is estimated to occur n sampling times ahead is absorbed in advance by correcting the position command Pt (k) before (current). Such control is made. In this case, when controlling the hand portion of the robot body 1 to follow the set target trajectory L, the robot body is not allowed to always coincide with the position command as in feedforward control, but a time delay is allowed. Control is performed so that the position of the one hand portion is placed on the target trajectory L.

  Accordingly, it is possible to reduce the positional deviation from the target trajectory L of the hand portion of the robot main body 1 due to the response delay of each servo motor 11, and the actual position of the hand portion of the robot main body 1 to the target trajectory L with high accuracy. Can be followed. At this time, control is performed with a certain delay with respect to the midpoint target point for each sampling time, but unlike conventional feedforward control, vibration due to excessive follow-up to the command value can be prevented. In addition, unlike the case of performing an adjustment to correct the teaching point by obtaining an error between the teaching locus and the locus when the actual robot body 1 is operated in advance, there is no need for an adjustment operation by repeated operations. Of course.

  As described above, according to the present embodiment, the robot body 1 is controlled so as to follow the set target trajectory, and the actual deviation of the trajectory of the robot main body 1 with respect to the target trajectory L is effective. Can be reduced. As a result, when the workpiece is machined with a machining tool, the time on the target track L is not required to be so strict as in the case of a sealing agent application operation, a welding operation, etc. This is effective when following is more important. In particular, in the present embodiment, when calculating the correction value, the perpendicular vector dropped on the target trajectory L from the estimated position of the hand of the robot body 1 is used as the correction vector. A correction value can be obtained, which is more effective.

  In the above embodiment, the perpendicular vector dropped on the target trajectory L from the estimated position of the hand of the robot body 1 is used as the correction value. It is also possible to use a vector directed to a point as a correction value. In addition, for example, various modifications can be made to the hardware configuration of the robot body, and the present invention is not limited to the above-described embodiments, and can be implemented with appropriate modifications without departing from the scope of the invention. Is.

1 is a control block diagram of position command correction according to an embodiment of the present invention. Flow chart showing the procedure of position command correction Diagram showing how the target trajectory and the actual trajectory shift Block diagram schematically showing the electrical configuration of the robot controller A perspective view schematically showing the appearance of the robot

Explanation of symbols

In the drawings, 1 is a robot body, 2 is a robot controller (control device), 11 is a servo motor, 12 is a rotary encoder, 13 is a CPU, 17 is a drive circuit, 18 is a position detection circuit, and L is a target trajectory.

Claims (4)

A robot control device that controls a robot's hand to follow a set target trajectory,
Position command output means for outputting a position command so as to move the robot's hand part to a target point on the way of subdividing the target trajectory for each sampling time;
A control unit for controlling each axis of the robot based on the position command;
Estimating means for estimating the position of the hand portion of the robot a predetermined sampling time from the present time;
Correction value calculating means for calculating a correction value of the position command by obtaining a deviation between the position estimated by the estimating means and the target trajectory;
An apparatus for controlling a robot, comprising: correction means for correcting a current position command based on a correction value calculated by the correction value calculation means.   The robot control apparatus according to claim 1, wherein the correction value calculation unit calculates a correction value from the position estimated by the estimation unit, using a perpendicular vector dropped on the target trajectory as a correction vector. By outputting a position command so as to move the hand of the robot to a target point halfway subdividing the set target trajectory at each sampling time, and controlling each axis of the robot based on the position command, A control method for a robot for controlling a hand portion of the robot to follow the target trajectory,
Estimate the position of the hand part of the robot ahead of the predetermined sampling time from the present,
Obtaining a deviation between the estimated position and the target trajectory and calculating a correction value of the position command;
A method for controlling a robot, wherein the current position command is corrected by the calculated correction value. 4. The robot control method according to claim 3, wherein the correction value is calculated from the estimated position of the hand of the robot as a correction vector using a perpendicular vector dropped on the target trajectory.

Images